شماره راهنما :
1189 دكتري
پديد آورنده :
كريمي، مريم
عنوان :
بهبود كارآيي روش‌هاي محاسباتي ارزيابي كيفيت ادراكي تصوير
گرايش تحصيلي :
هوش مصنوعي
محل تحصيل :
اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر
صفحه شمار :
بيست و يك، [162]ص.: مصور، جدول، نمودار
استاد راهنما :
شادرخ سماوي
توصيفگر ها :
ارزيابي كيفيت چشمي , ارزيابي كيفيت محاسباتي , كيفيت تجربي , كيفيت ادراكي , تصاوير بازنگار , تصاوير استريو , تركيب دوچشمي , آمارگان مكاني , بازنمايي تُنُك
استاد داور :
محمدرضا احمدزاده، الهام محمودزاده
تاريخ ورود اطلاعات :
1397/03/07
رشته تحصيلي :
برق و كامپيوتر
دانشكده :
مهندسي برق و كامپيوتر
كد ايرانداك :
ID1189 دكتري
چكيده انگليسي :
152 Improvement of Perceptual Objective Image Quality Assessment Methods Maryam Karimi Maryam karimi@ec iut ac ir 23 4 2018 Department of Electrical and Computer Engineering Isfahan University of Technology 84156 83111 Isfahan IranSupervisor Prof Shadrokh Samavi samavi96@cc iut ac irAdvisor Dr Nader Karimi nader karimi@cc iut ac ir Department of Electrical and Computer Engineering Isfahan University of Technology 84156 83111 Iran AbstractPerceptual quality assessment is important for testing optimizing and inspecting relevant multimediasystems services and algorithms Current objective scores are not close enough to subjective ones Thepurpose of this research is to use models based on the human visual system to promote objective qualityassessment systems In this regard first a binocular combination model for combining stereo pair images isproposed Then two no reference and general purpose objective quality assessment models are presentedbased on the feature extraction from these combined images The first model uses the spatial domain featuresof combined images to train a stacked generalizer The second model is based on sparse representation ofcombined images using pristine image dictionaries These methods have increased the best correlations withsubjective scores by up to 1 and 5 respectively In the next step two full reference methods for qualityevaluation of retargeted images are provided The first method provides three groups of features to coverdifferent types of retargeting distortions The second method introduces a pyramidal integration model fordifferent types of features unlike most of the existing methods that just use global pooling The proposedquality assessment methods provide rank correlation of up to 5 higher than the other methods Despite thesame efficiency the second method is much faster than the first one and is fully applicable for realapplications Key words subjective quality assessment Objective quality assessment Perceptualquality Stereo images Retargeted images Binocular combination Spatial natural scenestatistics Sparse representation IntroductionQuality of images degrades due to image processing applications such as compression retargeting and super resolution or by going through communication channels This hascreated a significant need for image quality assessment IQA at the end user side It isinconvenient costly and time consuming to have all video and images subjectivelyassessed Moreover subjective results are vulnerable to circumstances and individualcharacteristics of viewers Hence lots of research has been done to design objective imagequality assessment approaches based on the availability of reference image IQA methodsare divided into three major categories full reference FR reduced reference RR and noreference NR Stereo Image Quality AssessmentDue to the increasing popularity of three dimensional images efficient and fast 3D IQAmethods which are highly correlated with subjective mean opinion scores MOS aredesired for quality protection and improvement of 3D imaging applications The first 3DIQA category use depth in addition to 2D information of stereo pairs The second category
استاد راهنما :
شادرخ سماوي
استاد داور :
محمدرضا احمدزاده، الهام محمودزاده
